We propose to continue the development and maintenance of the LGXSM Recombinant Inbred (Rl) strain set and the associated Advanced Intercross (Al) Line formed by the intercross of inbred mouse strains LG/J and SM/J. The RIs are a tool for preliminary gene mapping, after which the Al Line can be used for finemapping quantitative trait loci to sub-cM regions. Under prior support we have developed a set of 18 Rl lines formed from the intercross of LG/J and SM/J and brought the Al line to the 32nd generation of random mating. The Rl lines have been scored for over 500 polymorphic microsatellite markers and for 4290 polymorphic SNPs, a polymorphic SNPs every 600 kb. Embryos have been cryopreserved for each Rl lines. These lines and other populations formed from the intercross of LG/J and SM/J have allowed the mapping of a large array of phenotypes relevant to disease processes of interest across the NIH Institutes. Here, we propose to enhance the Rl strain set by developing another 50 Rl strains using the present Al line as the source population. The addition of these strains will greatly increase the power of the Rl strain set in gene mapping studies. The new strains will be even more useful than the earlier set in that they will have accumulated 16 times the amount of recombination expected in Rl lines formed from a F2 generation. Furthermore, we will be able to use our prior breeding experience to increase strain survival to 50% in the production of new strains by selecting against specific loci that, singly and in combination, result in strain loss. The new strains will be genetically characterized for 500 microsatellites and 4290 SNPs and archived by embryonic cryopreservation. Strain genotypes and phenotypes will be made available through our web site, the Mouse Phenome Project, and the WebQTL database. In our earlier breeding, we discovered strong selection against the agouti alleles carried by SM/J, especially in the presence of the wild-type tyrosinase allele from that strain. We will examine the genetic basis for strain loss by sequencing the structural genes at these two loci and closely linked loci. We will produce 2ble congenic strains that are SM/J across the genome, except at the agouti and tyrosinase regions. Backcrossing with SM/J will be used to narrow the QTL interval for strain loss and to produce a new strain, carrying the unusual phenotypes of SM/J without requiring enforced heterozygosity at the agouti locus.